In: 2004 International conference on cyberworlds, pp 306–311. Bastanfard A, Takahashi H, Nakajima M (2004) Toward e-appearance of human face and hair by age, expression and rejuvenation. Bastanfard A Bastanfard O Takahashi H Nakajima M Toward anthropometrics simulation of face rejuvenation and skin cosmetic: research articles Comput Animat Virtual Worlds 2004 15 3–4 347 352 10.1002/cav.38 Google Scholar Digital Library Baek J Yoo Y Bae S Generative adversarial ensemble learning for face forensics IEEE Access 2020 8 45421 45431 10.1109/ACCESS.2020.2968612 Google Scholar Amirkhani D, Bastanfard A (2021) An objective method to evaluate exemplar-based inpainted images quality using jaccard index. In: 2019 IEEE/CVF International conference on computer vision workshop (ICCVW), pp 1205–1207. Amerini I, Galteri L, Caldelli R, Del Bimbo A (2019) Deepfake video detection through optical flow based cnn. In addition, the effectiveness of the proposed method is evaluated and compared with state-of-the art methods in terms of both, number of trainable parameters and binary detection accuracy. The popular FaceForensic++ dataset is employed to train and test the proposed method. Further, to understand the detection process of 3DCNN, activation maps are studied in detail. The proposed network learns spatio-temporal features to detect forged videos. The proposed 3DCNN is a five layered architecture with only 2.69M trainable parameters. Therefore, we propose a light weight 3D convolutional neural networks (3DCNN) to detect popular facial forgeries namely, DeepFakes, Face2Face and FaceSwap. Highly accurate facial forgery detectors will be incompatible with limited computation and storage space devices like smartphones, tablets, PCs due to their deep complex architectures. Detecting facial forgery in real time application such as video conferencing, netbanking, verification of identity etc.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |